MacPro6,1 2013 Black Can, 2 AMD D700 GPUS, using all TB2 ports for storage and v...
 
Notifications
Clear all

MacPro6,1 2013 Black Can, 2 AMD D700 GPUS, using all TB2 ports for storage and video, TB2 port1 to egpu, works in OSX, not linux.. PCI issue  

 of  2
  RSS

Jerry Normandin
(@jerry_normandin)
Active Member
Joined: 4 years ago
 

Hi,

      I have a MacPro6,1 black can desktop.   My I used the automate_egpu.sh script to setup the Akitio Node w/GTX 970 in Macs.. works great.

I figured it was time to beef up my Linux Bare metal environment so I picked up another 480gb SSD.. installed Ubuntu 17.10 Artful Ardvark.   I got my environment all built up..plugged in my eGPU and npticed this in dmesg:

[ 120.342191] pci 0000:19:00.0: BAR 1: no space for [mem size 0x10000000 64bit pref] [ 120.342195] pci 0000:19:00.0: BAR 1: failed to assign [mem size 0x10000000 64 ]

BAR 1 is for the Akitio Node.   I knew I wasn't going to be successful  but I proceeded along anyway building the latest and greatest NVIDIA driver.

As expected here's the error I got:

[ 252.499915] nvidia: module license 'NVIDIA' taints kernel.
[ 252.499918] Disabling lock debugging due to kernel taint
[ 252.516292] nvidia-nvlink: Nvlink Core is being initialized, major device num
ber 238
[ 252.516735] nvidia 0000:19:00.0: enabling device (0000 -> 0003)
[ 252.516839] NVRM: This PCI I/O region assigned to your NVIDIA device is inval
id:
NVRM: BAR1 is 0M @ 0x0 (PCI:0000:19:00.0)
[ 252.516840] NVRM: The system BIOS may have misconfigured your GPU.
[ 252.516847] nvidia: probe of 0000:19:00.0 failed with error -1
[ 252.516874] NVRM: The NVIDIA probe routine failed for 1 device(s).
[ 252.516875] NVRM: None of the NVIDIA graphics adapters were initialized!
[ 252.517019] nvidia-nvlink: Unregistered the Nvlink Core, major device number

19:00.0 VGA compatible controller: NVIDIA Corporation GM204 [GeForce GTX 970] (rev a1)
19:00.1 Audio device: NVIDIA Corporation GM204 High Definition Audio Controller (rev a1)

 

Here's my Nodes' entry on thunderbolt:

/sys/bus/thunderbolt/devices/0-1

root@bart-linux:/sys/bus/thunderbolt/devices/0-1# cat device_name
AKiTiO Node

oot@bart-linux:/sys/bus/thunderbolt/devices/0-1# cat authorized
1

I have removed the pci root devices and rscanned the bus like this:

root@bart-linux:/sys/bus/thunderbolt/devices/0-1/nvm_active0/power# lspci | grep -i pci | grep -i bridge | grep -i express
00:01.0 PCI bridge: Intel Corporation Xeon E7 v2/Xeon E5 v2/Core i7 PCI Express Root Port 1a (rev 04)
00:02.0 PCI bridge: Intel Corporation Xeon E7 v2/Xeon E5 v2/Core i7 PCI Express Root Port 2a (rev 04)
00:03.0 PCI bridge: Intel Corporation Xeon E7 v2/Xeon E5 v2/Core i7 PCI Express Root Port 3a (rev 04)
00:11.0 PCI bridge: Intel Corporation C600/X79 series chipset PCI Express Virtual Root Port (rev 06)
00:1c.0 PCI bridge: Intel Corporation C600/X79 series chipset PCI Express Root Port 1 (rev b6)
00:1c.1 PCI bridge: Intel Corporation C600/X79 series chipset PCI Express Root Port 2 (rev b6)
00:1c.2 PCI bridge: Intel Corporation C600/X79 series chipset PCI Express Root Port 3 (rev b6)
00:1c.4 PCI bridge: Intel Corporation C600/X79 series chipset PCI Express Root Port 5 (rev b6)

echo 1 > /sys/bus/pci/devices/0000\:00\:1c.0/remove

echo 1 > /sys/bus/pci/devices/0000\:00\:1c.1/remove

echo 1 > /sys/bus/pci/devices/0000\:00\:1c.2/remove

echo 1 > /sys/bus/pci/devices/0000\:00\:1c.4/remove

echo 1 > /sys/bus/pci/rescan

after attempting to install the kernel I get:

[ 5150.540735] NVRM: This PCI I/O region assigned to your NVIDIA device is invalid:
NVRM: BAR1 is 0M @ 0x0 (PCI:0000:19:00.0)
[ 5150.540738] NVRM: The system BIOS may have misconfigured your GPU.
[ 5150.540745] nvidia: probe of 0000:19:00.0 failed with error -1
[ 5150.540774] nvidia-nvlink: Nvlink Core is being initialized, major device number 239
[ 5150.540811] NVRM: The NVIDIA probe routine failed for 1 device(s).
[ 5150.540812] NVRM: None of the NVIDIA graphics adapters were initialized!
[ 5150.540814] nvidia-nvlink: Unregistered the Nvlink Core, major device number 239

 

Do I need to Reserve memory space for PCI allocation ?  I've been digging around looking for an PCI compile option.  I did try pci=realloc as a grub boot option and that did not work. 

 

Any help would be appreciated!

 

Thanks

 

To do: Create my signature with system and expected eGPU configuration information to give context to my posts. I have no builds.

.

ReplyQuote
Flint Ironstag
(@flint-ironstag)
Estimable Member
Joined: 5 years ago
 

Have you seen this?

http://pocketnix.org/posts/eGPUs%20under%20Linux%3A%20an%20advanced%20guide

Author is here, too- original thread for reference:

https://egpu.io/forums/thunderbolt-linux-setup/egpus-under-linux-an-advanced-guide/#post-19334

MP 6,1 | 4c | d700
MP 6,1 | 6c | d500


ReplyQuote
nando4
(@nando4)
Noble Member Admin
Joined: 5 years ago
 

Please read the Intro at https://egpu.io/forums/pc-setup/fix-dsdt-override-to-correct-error-12/ and consider using the Linux kernel 'pci=noCRS' option to allow 64-bit allocation of PCI device resources.

eGPU Setup 1.35    •    eGPU Port Bandwidth Reference Table

 
2015 15" Dell Precision 7510 (Q M1000M) [6th,4C,H] + GTX 1080 Ti @32Gbps-M.2 (ADT-Link R43SG) + Win10 1803 // compares M.2 vs TB3 performance inc unoptimized H-CPU BIOS [build link]  


ReplyQuote
Jerry Normandin
(@jerry_normandin)
Active Member
Joined: 4 years ago
 

Thanks both of you for the quick replay.   Yes.. the first thing I did was read all the documentation.  I should have included that in my original post.  I updated grub with

pci=noCRS.      It made my MacPro perform poorly.    I managed to log in and run dmesg to see what was going on.  There was a very high count of callbacks for every device kernel module communicating with a device on the pci bus.   I tried it again, this time with the eGPU disconnected.... again the MacPro was extremely slow.  I am talking 2 minutes to log in, and additional 3 minutes to open a terminal window and type dmesg.  Another minute to type reboot and wait for the reboot.

Resources are tight because of the Dual AMD D700 GPS, and I am using just about all the thunderbolt ports.  2 jbods,  2 monitors,  and the eGPU on port 1.   I have port 2 open.   The built in USB3 controller is on the PCIe bus too.

I looked the automate-eGPU.sh script.   The nvram write appears to be an authentication key.. we don't need to do that.  It appears that when automate_eGPU.sh is run it enables IOPCITunneling for the device.   Has anyone here been successful installing an eGPU in a Linux Environment running on a MacPro (late 2013) black can desktop with dual AMD D700 gpus ?  

 

To do: Create my signature with system and expected eGPU configuration information to give context to my posts. I have no builds.

.

ReplyQuote
Jerry Normandin
(@jerry_normandin)
Active Member
Joined: 4 years ago
 

I wanted to update this thread.  Almost working.   

[ 6.698736] nvidia: module license 'NVIDIA' taints kernel.
[ 6.698737] Disabling lock debugging due to kernel taint
[ 6.705722] nvidia-nvlink: Nvlink Core is being initialized, major device number 240
[ 6.705897] nvidia 0000:19:00.0: enabling device (0000 -> 0002)
[ 6.705991] nvidia 0000:19:00.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=none
[ 6.706060] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 384.90 Tue Sep 19 19:17:35 PDT 2017 (using threaded interrupts)
[ 6.716837] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 384.90 Tue Sep 19 17:05:19 PDT 2017
[ 6.718623] [drm] ring test on 5 succeeded in 2 usecs
[ 6.718626] [drm] UVD initialized successfully.

[ 8.700850] [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:19:00.0 on minor 2

 

erryn@bart-linux:~/Desktop$ nvidia-smi
No devices were found

[ 394.892201] NVRM: RmInitAdapter failed! (0x26:0xffff:1113)
[ 394.892266] NVRM: rm_init_adapter failed for device bearing minor number 0

jerryn@bart-linux:~/Desktop$ ls -l /dev/nvidia*
crw-rw-rw- 1 root root 195, 0 Oct 13 20:09 /dev/nvidia0
crw-rw-rw- 1 root root 195, 255 Oct 13 20:09 /dev/nvidiactl

 

Is anyone using NVIDIA-Linux-x86_64-384.90  ?  This should be working.  

To do: Create my signature with system and expected eGPU configuration information to give context to my posts. I have no builds.

.

itsage liked
ReplyQuote
jacooba
(@jacooba)
Active Member
Joined: 3 years ago
 
Hi there, I'm new here and I could really use some help! 
 
I have Ubuntu 16.04 LTS installed on my MBP (15 inch, touchbar, late 2016 i believe). I am trying to get my gamebox aurous with a gtx 1080 working. I have it working on windows, but I need it to work with python on linux for my work.
 
Right now, when I try to insert the driver I get "insmod: ERROR: could not insert module /lib/modules/4.13.0-45-generic/updates/dkms/nvidia_384.ko: No such device". (also "modprobe nvidia" gives "modprobe: FATAL: Module nvidia not found in directory /lib/modules/4.13.0-45-generic")
 
In addition, dmesg says "NVRM: BAR1 is 0M @ 0x0 (PCI:0000:0a:00.0)
[   85.546913] NVRM: The system BIOS may have misconfigured your GPU."
 
I have tried all of the following:
1. "echo 1 > /sys/bus/pci/devices/0000\:00\:1c.0/remove
2. and echo 1 > /sys/bus/pci/rescan
3. adding pci=realloc to my kernel parameters
4. adding pci=nocrs
5. installing nvidia 384 driver
6. installing  nvidia-modprobe
7. installing nvidia 384 from a runfile (which i couldnt get to work for various reasons)
8. installing the driver from the ppa repository
9. installing through the "additional drivers" page in settings.
 
 
 
 
 
 
If it is any help, the output of dmesg and lspci are bellow:
 
 
 
 
00:00.0 Host bridge: Intel Corporation Sky Lake Host Bridge/DRAM Registers (rev 07)
00:01.0 PCI bridge: Intel Corporation Sky Lake PCIe Controller (x16) (rev 07)
00:01.1 PCI bridge: Intel Corporation Sky Lake PCIe Controller (x8) (rev 07)
00:01.2 PCI bridge: Intel Corporation Sky Lake PCIe Controller (x4) (rev 07)
00:14.0 USB controller: Intel Corporation Sunrise Point-H USB 3.0 xHCI Controller (rev 31)
00:15.0 Signal processing controller: Intel Corporation Sunrise Point-H LPSS I2C Controller #0 (rev 31)
00:16.0 Communication controller: Intel Corporation Sunrise Point-H CSME HECI #1 (rev 31)
00:19.0 Signal processing controller: Intel Corporation Sunrise Point-H LPSS UART Controller #2 (rev 31)
00:1b.0 PCI bridge: Intel Corporation Sunrise Point-H PCI Root Port #17 (rev f1)
00:1c.0 PCI bridge: Intel Corporation Sunrise Point-H PCI Express Root Port #1 (rev f1)
00:1e.0 Signal processing controller: Intel Corporation Sunrise Point-H LPSS UART #0 (rev 31)
00:1e.1 Signal processing controller: Intel Corporation Sunrise Point-H LPSS UART #1 (rev 31)
00:1e.2 Signal processing controller: Intel Corporation Sunrise Point-H LPSS SPI #0 (rev 31)
00:1e.3 Signal processing controller: Intel Corporation Sunrise Point-H LPSS SPI #1 (rev 31)
00:1f.0 ISA bridge: Intel Corporation Sunrise Point-H LPC Controller (rev 31)
00:1f.2 Memory controller: Intel Corporation Sunrise Point-H PMC (rev 31)
00:1f.3 Audio device: Intel Corporation Sunrise Point-H HD Audio (rev 31)
00:1f.4 SMBus: Intel Corporation Sunrise Point-H SMBus (rev 31)
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Device 67ef (rev c0)
01:00.1 Audio device: Advanced Micro Devices, Inc. [AMD/ATI] Device aae0
02:00.0 Non-Volatile memory controller: Samsung Electronics Co Ltd Device a804
03:00.0 Network controller: Broadcom Corporation BCM43602 802.11ac Wireless LAN SoC (rev 02)
04:00.0 PCI bridge: Intel Corporation DSL6540 Thunderbolt [Alpine Ridge] (rev 02)
05:00.0 PCI bridge: Intel Corporation Device 15d3 (rev 02)
05:01.0 PCI bridge: Intel Corporation Device 15d3 (rev 02)
05:02.0 PCI bridge: Intel Corporation Device 15d3 (rev 02)
05:04.0 PCI bridge: Intel Corporation Device 15d3 (rev 02)
06:00.0 System peripheral: Intel Corporation Device 15d2 (rev 02)
07:00.0 USB controller: Intel Corporation Device 15d4 (rev 02)
08:00.0 PCI bridge: Intel Corporation DSL6540 Thunderbolt [Alpine Ridge]
09:01.0 PCI bridge: Intel Corporation DSL6540 Thunderbolt [Alpine Ridge]
09:04.0 PCI bridge: Intel Corporation DSL6540 Thunderbolt [Alpine Ridge]
0a:00.0 VGA compatible controller: NVIDIA Corporation Device 1b80 (rev a1)
0a:00.1 Audio device: NVIDIA Corporation Device 10f0 (rev a1)
0b:00.0 USB controller: Intel Corporation DSL6540 USB 3.1 Controller [Alpine Ridge]
7a:00.0 PCI bridge: Intel Corporation DSL6540 Thunderbolt [Alpine Ridge] (rev 02)
7b:00.0 PCI bridge: Intel Corporation Device 15d3 (rev 02)
7b:01.0 PCI bridge: Intel Corporation Device 15d3 (rev 02)
7b:02.0 PCI bridge: Intel Corporation Device 15d3 (rev 02)
7b:04.0 PCI bridge: Intel Corporation Device 15d3 (rev 02)
7c:00.0 System peripheral: Intel Corporation Device 15d2 (rev 02)
7d:00.0 USB controller: Intel Corporation Device 15d4 (rev 02)
 
 
 
 
 
 
 
[   63.948786] nvidia-nvlink: Nvlink Core is being initialized, major device number 240
[   63.949134] NVRM: This PCI I/O region assigned to your NVIDIA device is invalid:
               NVRM: BAR1 is 0M @ 0x0 (PCI:0000:0a:00.0)
[   63.949135] NVRM: The system BIOS may have misconfigured your GPU.
[   63.949139] nvidia: probe of 0000:0a:00.0 failed with error -1
[   63.949167] NVRM: The NVIDIA probe routine failed for 1 device(s).
[   63.949167] NVRM: None of the NVIDIA graphics adapters were initialized!
[   63.949326] nvidia-nvlink: Unregistered the Nvlink Core, major device number 240
[   76.475919] nvidia-nvlink: Nvlink Core is being initialized, major device number 240
[   76.476146] NVRM: This PCI I/O region assigned to your NVIDIA device is invalid:
               NVRM: BAR1 is 0M @ 0x0 (PCI:0000:0a:00.0)
[   76.476147] NVRM: The system BIOS may have misconfigured your GPU.
[   76.476149] nvidia: probe of 0000:0a:00.0 failed with error -1
[   76.476164] NVRM: The NVIDIA probe routine failed for 1 device(s).
[   76.476164] NVRM: None of the NVIDIA graphics adapters were initialized!
[   76.476242] nvidia-nvlink: Unregistered the Nvlink Core, major device number 240
[   83.730141] pcieport 0000:09:04.0: bridge window [io  0x1000-0x0fff] to [bus 0b] add_size 1000
[   83.730157] pcieport 0000:08:00.0: bridge window [io  0x1000-0x0fff] to [bus 09-0b] add_size 1000
[   83.730164] pcieport 0000:05:01.0: bridge window [io  0x1000-0x0fff] to [bus 08-40] add_size 1000
[   83.730166] pcieport 0000:05:01.0: bridge window [mem 0x00100000-0x000fffff 64bit pref] to [bus 08-40] add_size 200000 add_align 100000
[   83.730202] pcieport 0000:7b:02.0: bridge window [mem 0x00100000-0x000fffff 64bit pref] to [bus 7d] add_size 200000 add_align 100000
[   83.730249] pcieport 0000:05:01.0: BAR 15: assigned [mem 0xc0200000-0xc03fffff 64bit pref]
[   83.730251] pcieport 0000:05:01.0: BAR 13: assigned [io  0x6000-0x6fff]
[   83.730254] pcieport 0000:08:00.0: BAR 13: assigned [io  0x6000-0x6fff]
[   83.730256] pcieport 0000:09:04.0: BAR 13: assigned [io  0x6000-0x6fff]
[   83.730261] pcieport 0000:7b:02.0: BAR 15: no space for [mem size 0x00200000 64bit pref]
[   83.730263] pcieport 0000:7b:02.0: BAR 15: failed to assign [mem size 0x00200000 64bit pref]
[   83.730265] pcieport 0000:7b:02.0: BAR 15: no space for [mem size 0x00200000 64bit pref]
[   83.730266] pcieport 0000:7b:02.0: BAR 15: failed to assign [mem size 0x00200000 64bit pref]
[   85.546651] nvidia-nvlink: Nvlink Core is being initialized, major device number 240
[   85.546912] NVRM: This PCI I/O region assigned to your NVIDIA device is invalid:
               NVRM: BAR1 is 0M @ 0x0 (PCI:0000:0a:00.0)
[   85.546913] NVRM: The system BIOS may have misconfigured your GPU.
[   85.546916] nvidia: probe of 0000:0a:00.0 failed with error -1
[   85.546928] NVRM: The NVIDIA probe routine failed for 1 device(s).
[   85.546929] NVRM: None of the NVIDIA graphics adapters were initialized!
[   85.547017] nvidia-nvlink: Unregistered the Nvlink Core, major device number 240
[  161.316492] IPv6: ADDRCONF(NETDEV_UP): wlp3s0: link is not ready
[  161.352795] brcmfmac: brcmf_inetaddr_changed: fail to get arp ip table err:-23
[  162.840829] IPv6: ADDRCONF(NETDEV_CHANGE): wlp3s0: link becomes ready
[  172.415719] brcmfmac: brcmf_inetaddr_changed: fail to get arp ip table err:-23
[  177.635318] brcmfmac: brcmf_inetaddr_changed: fail to get arp ip table err:-23
[  182.277551] brcmfmac: brcmf_inetaddr_changed: fail to get arp ip table err:-23
[ 1026.005493] nvidia-nvlink: Nvlink Core is being initialized, major device number 240
[ 1026.005735] NVRM: This PCI I/O region assigned to your NVIDIA device is invalid:
               NVRM: BAR1 is 0M @ 0x0 (PCI:0000:0a:00.0)
[ 1026.005735] NVRM: The system BIOS may have misconfigured your GPU.
[ 1026.005738] nvidia: probe of 0000:0a:00.0 failed with error -1
[ 1026.005753] NVRM: The NVIDIA probe routine failed for 1 device(s).
[ 1026.005754] NVRM: None of the NVIDIA graphics adapters were initialized!
[ 1026.005845] nvidia-nvlink: Unregistered the Nvlink Core, major device number 240
 
 
 
 

To do: Create my signature with system and expected eGPU configuration information to give context to my posts. I have no builds.

.

ReplyQuote
karatekid430
(@karatekid430)
Estimable Member
Joined: 4 years ago
 

@jacooba insmod command wants the path to the .ko file, I think. Use modprobe command which happily accepts the module name. Technically that issue has nothing to do with external graphics in particular.

Also, I would advise to use in-tree drivers from the kernel - they are open-source and do not interfere with secure boot. Using proprietary drivers often shackles you to a particular kernel version, with no ability to upgrade if you need features from newer kernels.

In my experience, PCI=realloc and maybe PCI=nocrs (but definitely the first) can make Thunderbolt behave strangely, in some circumstances. They are excellent options for messing with Thunderbolt and trying to figure out the underlying mechanisms of operation. But unfortunately, until they free responsibility of Thunderbolt from the firmware / BIOS / ACPI completely, these options are probably better not used for production use.

Also, I only just noticed you are only using 4.13 kernel - which only just introduced Thunderbolt support. I would advise to take at the very least 4.16 (but 4.17 is highly recommended). Assuming Debian / Ubuntu, you can install from  http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.17.1/ and "dpkg -i" command as root to install them. Want headers (generic, all), linux-image-generic, linux-modules-generic, so total of four. If they are the only debs in a directory, "dpkg -i *.deb" should work as root.

Good luck!

To do: Create my signature with system and expected eGPU configuration information to give context to my posts. I have no builds.

.

ReplyQuote
karatekid430
(@karatekid430)
Estimable Member
Joined: 4 years ago
 

@jacooba I just remembered a couple of things:

1. Passing PCI=rom might be of more use than the other ones (and make sure you are *not* passing PCI=nobar). Here, have an rst:  https://www.kernel.org/doc/html/v4.17/admin-guide/kernel-parameters.html
2. The 2016 MacBook has some funky problem with firmware that causes Thunderbolt issues. There is a patch / workaround, that apparently Apple have applied officially in the 2017 model. I have not heard if they have updated the firmware in the 2016 model with the same patch. Something to do with DSDT.

But we cannot be sure until you get an up to date kernel and get rid of the proprietary Nvidia drivers - as the problems might be only because of those. Report back then. Also "lspci -vt" would be nice (make sure both Thunderbolt controllers are awake at the time, i.e. ports on both sides of laptop). Thanks.

To do: Create my signature with system and expected eGPU configuration information to give context to my posts. I have no builds.

.

ReplyQuote
jacooba
(@jacooba)
Active Member
Joined: 3 years ago
 
Thank you so much for the help @karatekid430 !
 
By "in-tree" drivers, do you mean the PPA repo?
 
As for modprobe, it did not find the module (output in above post), which is why i gave the path to insmod.
 
i just ran dpkg -i *.deb on the packages you mentioned, but I'm not sure if it worked correcty. I got a couple concering messages and apt doesn't work anymore. What should I do?:
 
from the terminal:
Setting up linux-headers-4.17.1-041701 (4.17.1-041701.201806111730) ...
dpkg: dependency problems prevent configuration of linux-headers-4.17.1-041701-generic:
 linux-headers-4.17.1-041701-generic depends on libssl1.1 (>= 1.1.0); however:
  Package libssl1.1 is not installed.
 
dpkg: error processing package linux-headers-4.17.1-041701-generic (--install):
 dependency problems - leaving unconfigured
 
Processing triggers for linux-image-unsigned-4.17.1-041701-generic (4.17.1-041701.201806111730) ...
/etc/kernel/postinst.d/dkms:
: Unable to find an initial ram disk that I know how to handle.
Will not try to make an initrd.
/etc/kernel/postinst.d/initramfs-tools:
update-initramfs: Generating /boot/initrd.img-4.17.1-041701-generic
W: Possible missing firmware /lib/firmware/amdgpu/vega12_gpu_info.bin for module amdgpu
W: Possible missing firmware /lib/firmware/amdgpu/vega12_asd.bin for module amdgpu
W: Possible missing firmware /lib/firmware/amdgpu/vega12_sos.bin for module amdgpu
 
Errors were encountered while processing:
 linux-headers-4.17.1-041701-generic
 
 
I also now have a red circle in the top bar on my desktop. It says "An error occured. Please run Package Manager or apt-get to see what is wrong. The error message was: 'Error: broken count >0'. This usually means that your installed packages have unmet dependencies."
There is an "update" option that pulls up the software updater. I can press "partial update" or" continue". Should I do one of these?
 
I ran " sudo apt-get install --fix-broken". This "The following packages will be REMOVED:
  linux-headers-4.17.1-041701-generic".
Should I do that?
 
(Side note, my screen brightness no longer seems to be adjustable after the kernel update, but I'm guessing I just have to reinstall whatever patch I had installed for that on the macbook.)

To do: Create my signature with system and expected eGPU configuration information to give context to my posts. I have no builds.

.

ReplyQuote
karatekid430
(@karatekid430)
Estimable Member
Joined: 4 years ago
 

@jacooba By in-tree drivers, I mean the ones that are built right into the Linux kernel (the source code is part of the Linux kernel). You should not need to install anything for the GPU to work. No PPA or anything.

Yes, you will probably have to remove the Linux 4.17 packages to get back to a state with all dependencies satisfied so that your package manager will agree to install other things.
"dpkg -l | grep linux" should show you all the packages installed (notice -l for list not -i for install) and then you can do "sudo apt-get purge package1 package2 …. packageN" with all four of the 4.17 package names to remove them, if nothing else works.

"Package libssl1.1 is not installed."
This line is why it failed.

Once you have removed the 4.17 packages,
sudo apt-get install -y libssl1.1
And then try to install Linux 4.17 again.

Also would not hurt to do "sudo apt-get install git" and then "git clone  https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.gi t" and "cd linux-firmware" and "sudo make install" to make sure you are up to date. Unfortunately the linux-firmware is no longer part of the main Linux git, and is maintained separately (which means it can get out of date with the kernel). Unless you run Arch Linux (not advised until you are extremely competent at Linux) then chances are your distro does not provide the latest packages.

Try to purge all traces of the Nvidia drivers you downloaded to make sure they are not interfering. Try with the one inbuilt in the kernel (which should automatically load when GPU is approved). Only bother with the installable drivers (proprietary) if the in-built drivers are not adequate or do not have a special feature you need.

Are you aware of how to check your GPU is working? "sudo apt-get install mesa-utils"
"DRI_PRIME=1 glxinfo | grep -i renderer" and hopefully it says something about Nvidia, instead of Intel graphics.
Prepending DRI_PRIME=1 should make an application use the big GPU, instead of the internal.
If you want to make it permanent, system-wide, open /etc/environment file as root in an editor, add the line "DRI_PRIME=1" without quotes, save, reboot. To check it worked, open a terminal and do "echo $DRI_PRIME" and if it prints "1" then it worked. If you only want to do it for your local user then need to add the line in ~/.bashrc or ~/.profile or something (I cannot remember which is the best, but you can ask the internet for that one "add environment variable to profile Linux Ubuntu").

If you want to start over, you can back up your home directory (and anywhere else where files are installed), and preferably also back up anything on the MacOS just in case you accidentally erase that too. Then download the Ubuntu 18.04 ISO on Linux, find a USB drive with no information (will be erased), make sure the USB is unmounted (system might automatically mount it when you plug in). Useful command "lsblk" to see drives and mount points.
"sudo dd if=/path/to/Ubuntu.iso of=/dev/sdX" then "sync". Make sure /dev/sdX is definitely the path to your USB (you can easily overwrite your main hard disk if you type the wrong thing).

Then you can boot off that USB and install Ubuntu over the top, and start again. Or to get Ubuntu 18.04 upgrade without starting over, "sudo do-release-upgrade -d" "sudo apt-get update" "sudo apt-get dist-upgrade" "sudo apt-get upgrade" etc.

Or you can search the internet for more comprehensive guides. Or you might already know all this. If so, my apologies.

To do: Create my signature with system and expected eGPU configuration information to give context to my posts. I have no builds.

.

ReplyQuote
 of  2